R The Specificity of Attention Retraining in E P Traumatic Brain Injury O R T Louise Penkman Catherine A. Mateer

Attentional deficits are a very common complaint following that attention training tasks and intensity of treatment do vary traumatic brain injury (TBI). Concentration problems, together widely across studies. In addition, an emphasis on training of with poor memory, are reported to be among the most common attention skills is often only one component of a more post concussive cognitive complaints (Binder, 1986). It has been comprehensive rehabilitation program. The program may also suggested that many apparent failures of memory are due to include awareness training and/or psychotherapeutic support to attention dysfunction and that this problem contributes assist the individual in dealing with the emotional consequences substantially to difficulty with reintegration into independent of sustaining a brain injury and coping with changes in living and vocational settings (Sohlberg & Mateer, 1989). Brooks functioning. Therefore, it remains unclear whether attention and McKinlay (1987) reported that attentional impairment predicts training per se is beneficial (Goransen, 1997). return to work after head injury. For example, Palmese and Raskin (2000) administered attention There has been considerable clinical and research interest in training to three adults with mild traumatic brain injury. Following the potential for improving attention skills. Attention interventions their treatment, the three adults participated in six to seven typically involve administration of a series of cognitive exercises sessions of brain injury education and a focus on the application that place increasingly greater demands on the attention system. of cognitive abilities. This condition was considered to be a no- Research in this area has produced somewhat mixed results. treatment control. However, all three of the participants showed While many studies have provided support for the efficacy of some change after both the attention training condition and the attention training, (Sohlberg & Mateer, 1987; Neimann, Ruff & no-treatment control condition. It is difficult to ascertain if the Baser, 1990; Gray & Robertson, 1989; Gray, Robertson, Pentland differences in performance on neuropsychological tasks noted & Anderson, 1992; Sturm, Willmes, Orgass & Hartje, 1997; after the no-treatment control were due to carry-over effects of Cicerone et al., 2000; Sohlberg, McLaughlin, Pavese, Heidrich & the attention training or to non-specific change related to both Posner, 2000; Palmese & Raskin, 2000), other authors have found conditions because there was no counter-balancing for order. more equivocal results (Gauggel & Neimann, 1996; Park, Proulx Similarly, other studies have included a psychotherapeutic & Towers, 1999; Malec, Jones, Rao & Stubbs, 1984; Gansler & element as part of the intervention but have not evaluated its McCaffrey, 1991; Ponsford & Kinsella, 1988). impact on outcome (Park, Proulx, & Towers, 1999; Middleton et Park and Ingles (2001) conducted a meta-analysis of 30 studies al., 1991). that evaluated the efficacy of attention interventions. They In a large meta-analysis, Carney and colleagues (1999) reported concluded that retraining attention only produces a small, non- no treatment effect when cognitive rehabilitation was compared statistically significant improvement on outcome measures. with another form of treatment. They stated the need to test the However, their study sample included studies evaluating differential effects of general stimulation versus cognitive interventions for individuals with stroke and surgical lesions in rehabilitation. Similarly, Wilson (1993) has suggested asking addition to traumatic brain injury. Differential response for these questions that tease out the effects of particular rehabilitation groups to rehabilitative interventions may have impacted their procedures as they are applied to particular patient groups. results. The patients represented in their meta-analysis had A few studies do attempt to provide data relevant to this primarily severe injuries so their findings cannot be generalized issue. Ruff and colleagues (1989) conducted a randomized group to individuals with injuries of mild to moderate severity. The design study aimed at separating specific treatment effects from nature and duration of the interventions evaluated were highly those of stimulation and social support. In their study two groups variable. Outcome measures were also highly variable and were of 20 subjects received different treatments. Their control group analyzed in aggregate and this may also have contributed to the received sessions focusing on psychosocial adjustment, leisure, negative findings. In addition, the exclusion criteria described in activities of daily living, coping skills and health. The experimental this meta-analysis were somewhat unclear and excluded frequently group received computerized training in attention, memory, spatial cited studies in this area (e.g., Sohlberg & Mateer, 1987). integration, and problem solving. Both groups received 50 minutes A significant problem with this body of literature is the fact of group psychotherapy at the beginning of each day. At post- Louise Penkman, PhD., Alberta Children’s Hospital, Oncology Pro- test, both groups improved significantly on neuropsychological gram, 1820 Richmond Road, S. W., Calgary, Alberta, T2T 5C7 and measures in all areas except for spatial integration. The relative Catherine A. Mateer, Department of , University of Victoria, efficacy of the two treatments was not demonstrated but statistical Victoria, British Columbia, Canada. trends in the data suggested that the experimental group made

Spring 2004 - The Journal of Cognitive Rehabilitation 13 R E P O R T

larger gains in almost all areas. Their data suggest that the enriched further support for the efficacy of APT in improving attentional environment characteristic of the therapeutic milieu yields some functioning. The data also strongly suggest that non-specific or benefits, even in the absence of targeted, neuropsychologically psychosocial interventions that are often part of a rehabilitation based training. Their data also suggest that specific training may program do exert an effect and should not be ignored in studies have some benefit above and beyond what is provided by a examining the efficacy of cognitive rehabilitation or attention supportive therapeutic situation but this conclusion is tentative. retraining techniques. The primary goal of the current study was An additional difficulty with this study is the fact that all to determine if there is a specific effect for the administration of participants received psychotherapy and the impact of this attention training above and beyond a supportive, compensatory intervention on outcome was not considered. In a later paper, and adjustment-focused approach. Ruff and Niemann (1990) reported on the impact of cognitive A secondary goal of this study was to determine if remediation versus psychosocial day treatment on emotional and neurophysiological correlates of attention change after attention psychosocial variables. They predicted an increase in affective retraining, implicating the occurrence of underlying neuronal distress after participation in cognitive remediation and a decrease recovery/change. The P300 and N200 event related potentials after participation in the day program. Using the design described (ERPs) were recorded because of their association with cognitive above, they found that their hypotheses were not supported. processes. The P300 is considered a neurophysiologic index of Both groups demonstrated a reduction in symptoms of attention capacity in humans (Deacon-Elliott, Campbell, Suffield psychosocial distress. These findings underscore the importance & Proulx, 1987; Polich, 1998; Picton, 1992). The N200 is associated of considering separately the impact of cognitive rehabilitation with evaluation of stimulus information required for response and psychosocial interventions in outcome studies. selection (Gevins & Cutillo, 1971). The P300 has been shown to In a study focused specifically on attention training, Sohlberg, be both reduced in amplitude and delayed in latency in the TBI McLaughlin, Pavese, Heidrich and Posner (2000) compared population (Heinze, Munte, Gobiet, Niemann & Ruff, 1992; Attention Process Training (APT) (Sohlberg & Mateer, 1987) Papincolau, Levin, Eisenberg, Moore, Goethe & High, 1984; with a placebo condition of brain injury education using a within Campbell, Deacon-Elliot & Proulx, 1986; Rugg et al., 1988; Keren, and between groups design. Fourteen individuals with brain Ben-Dror, Stern, Goldberg & Groswasser, 1998). injury were divided into two groups and received 24 hours of Researchers have begun to employ these measures to evalu- individually administered APT over 10 weeks and 10 hours of ate brain changes following participation in attention retraining brain injury education over 10 weeks in counterbalanced order. with individuals with TBI. Baribeau, Ethier & Braun (1989) as- Dependent measures were administered at pre-test and after each sessed auditory ERPs as neurophysiological indices of selective experimental condition. Outcome measures included attention before and after an intensive, computer-dispensed cog- neuropsychological tests sensitive to attention dysfunction, nitive rehabilitation program. They utilized 21 individuals with cognitive marker tasks (tasks that have been studied by positron TBI and also employed a control group of 22 individuals with emission tomography (PET) and for which there are putative TBI matched for age, gender and education. They found the neuroanatomical correlates), self-report questionnaires and a neurophysiological indicators to be sensitive to treatment con- structured interview. Results indicated practice effects and dition; however, there were no changes in error rates or speed. improvement on cognitive tasks after participation in both They report reduced N100 latency, increased amplitude of the conditions. However, participants showed greater improvement P1-N1-P2 component and increased amplitude of the Nd for the on several neuropsychological tests following APT. Structured left ear only for their experimental group. The authors interpreted interview information was examined as changes in three areas: their results as suggesting no improvement in selective attention everyday functioning, psychosocial functioning and attention/ but as possibly indicating increased motivation, attentional ef- memory. The number of changes reported was greater after APT fort and improved tonal stimulus processing. than after brain injury education. However, more psychosocial Stone and Raskin (1996) used quantified EEG as a measure of improvements were noted by participants after brain injury efficacy in a rehabilitation study evaluating the efficacy of education and more memory/attention improvements were noted prospective memory training with two participants with TBI. after APT. These participants demonstrated an abnormal frontal distribution Taken together, the data from these studies provide evidence of alpha activity prior to training which reverted to a more typical, of differential outcome from different treatment approaches and posterior distribution after training. Raskin (1996) also measured

14 Spring 2004 - The Journal of Cognitive Rehabilitation P300 before and after prospective memory training. The participant did report a history of alcohol abuse. Participant 1 R participants demonstrated reduced P300 latency after training. suffered two significant TBIs. The first was due to a fall from a E Therefore, there is some preliminary evidence pointing to the horse and the second was a bicycle - motor vehicle accident. possibility of brain related changes, as measured by changes in Participants 2 & 3 were injured in motor vehicle accidents. P brain electrical activity, as a consequence of direct interventions Participant 4 was assaulted. Participants 2 & 4 showed evidence O and practice on specific cognitive skills. of depressive symptoms on the Rand Depression Scale. R In the present study it was predicted that both Participant characteristics are displayed in Table 1. T neuropsychological test scores and neurophysiological measures would show significantly more positive change after participation Design and Procedures in an attention training condition (using Attention Process Training materials-APT) than after participation in an Active A modified single case multiple baseline cross-over design Control condition (AC). With respect to specific ERP components, was utilized. Each participant underwent a six-week baseline it was predicted that the P300 and N200 components would show phase, six weeks participation in Attention Process Training (APT) differences after participation in the attention training condition. and six weeks participation in an Active Control (AC) condition. More specifically, it was predicted that a decrease in latency of Participants took part in the AC condition together in a group the both the P300 and N200 components would be observed. In format. Order of administration of the two conditions was addition, a decrease in the negativity of the N200 was predicted. counterbalanced (Participants 1 & 2 received APT first; It was predicted that P300 amplitude would also show change, Participants 3 & 4 received AC first). Frequency and intensity of but specific hypotheses about directionality of this change were the intervention was identical for the two conditions (twice per not offered because the P300 component is influenced by both week for one hour each session). task difficulty and level of certainty (Picton, 1992); therefore we Behavioral dependent measures were collected at two-week were unclear as to how this might influence our findings. intervals throughout the study for a total of nine data points per measure. Neurophysiological measures were recorded at two- Method week intervals during the baseline phase and were collected at the end of each treatment condition for a total of five recordings. Participants Testing was carried out by a research assistant not involved in the provision of the rehabilitation procedures. The therapist was Four adults with moderate to severe TBI participated in this the same for both conditions to control for therapist effect. study. They were selected from a pool of 21 available participants. Initially, seven participants meeting inclusion criteria were enrolled Active Control Condition in the study. However, two participants dropped out and one The Active Control condition was outlined in a manual format was not able to attend enough sessions to collect reliable data. and was designed to focus on emotional adjustment and Prior to study entry, participants took part in a clinical interview psychoeducation and to control for the effects of general and a brief neuropsychological assessment in order to document stimulation and therapeutic and social support. Different topics their cognitive difficulties. In addition they completed The relevant to coping with TBI were discussed for each session. Depression Scale developed by the Rand Corporation. It is a very short questionnaire that correlates reasonably well with the Table 1 Beck Depression Scale, r = .70, p < .05. It contains only items of Participant Characteristics an emotional nature and does not contain any of the vegetative or cognitive items that may be confounded with brain injury Gender Age TSI Neuroimaging LOC symptomatology. Participants selected were at least one year post injury and 1 Female 58 (1) 26 mos Subdural hematoma 9 days coma evidenced an injury of at least moderate severity as defined by (2) 40 yrs with midline shift, either a Glasgow Coma Scale (GCS) of 8 to 12, loss of skull fracture consciousness of 30 minutes or more, a period of post traumatic right temporal contusion amnesia (PTA) of at least one hour, or evidence of a brain injury documented by neuroimaging studies. Performance on the Paced 2 Male 22 45 mos n/a Brief with 3 Auditory Serial Addition Test (PASAT) (Gronwall, 1977) was at days PTA least 1.5 standard deviations below the mean. Because of the single case experimental design nature of this study individual 3 Male 34 12 yrs Right sided injury 8 wks coma variability was of interest. Therefore, participants were not with left hemiparesis excluded on the basis of seizures or previous TBIs. In addition, and blindness participants were not excluded on the basis of a prior history of substance abuse. Alcohol abuse is common in individuals who 4 Male 40 15 mos n/a 2 days with sustain traumatic brain injuries (Grafman & Salazar, 1987); significant therefore to exclude participants on this basis would not be PTA representative of the characteristics of the population of interest. However, participants were screened to ensure that they were Note: LOC = loss of consciousness, PTA = post-traumatic not currently abusing either alcohol or street drugs. One amnesia, TSI = time since injury

Spring 2004 - The Journal of Cognitive Rehabilitation 15 Topics included: education about brain injury, fatigue, headaches, For recording of ERPs, participants engaged in a visual and problem solving, relaxation training, use of compensations, and an auditory oddball task (detecting targets in a sequence of R cognitive restructuring. standard stimuli). For the auditory task they listened to a series E of tone beeps presented binaurally through earphones and were Attention Training Condition required to count the number of targets (designated as the higher P The attention retraining module consisted of the Attention tone). Standard or non-target tones were 2000 Hz and targets O Process Training (APT) materials (Sohlberg & Mateer, 1987). The were 1000 Hz. Both tones were presented at an 80 dB level. Targets R APT materials consist of hierarchically organized treatment tasks were presented 15% of the time (p=.l5) and non-targets were T for each of the five levels of attention outlined in Sohlberg and presented 85% of the time (p=. 85). This task is believed to tap Mateer’s componential model of attention: focused, sustained, working memory and attention processes. For the visual task, selective, alternating, and divided attention. The materials include participants were presented with blue triangles (non-targets, paper and pencil tasks; some exercises are presented on p=.85) and blue circles (targets, p=.l5) on a computer monitor. audiotape. Other activities include multi-tasking such as sorting They were required to press a button whenever a blue circle a deck of cards and counting at the same time. appeared on the screen. Reaction times were recorded for this task. Measures ERP Recording Measures of Attention and Working Memory Electrodes were placed on the scalp for 3 response sites (FZ, Participants completed two versions of a short computerized PZ, CZ - International 10-20 System). These sites are located on working memory task (Kerns, 1997; Owens, Morris, Sahakian, the midline and cover the frontal to parietal area. An electrode Polkey & Robbin, 1996). They were presented with a number of cap was used. Electrodes placed on the mastoid served as boxes and words on a computer screen. They were required to references. Electrode impedance was kept below ten K Ω . ERPs locate symbols hidden beneath the words or boxes. This involved were recorded using a Bio-Logic Brain Atlas (BioLogic Systems). remembering where the previously seen boxes were located The signal was amplified 50,000 times and filtered with a 60 Hz (spatial memory) and which words they had already chosen notch filter with a band pass of 1-30 Hz. Sampling rate was 200 Hz (verbal memory). Location of the symbols was randomly from 100 msec prior to stimulus onset to 924 msec post-stimulus. determined by the computer program. Four trials for each task An automatic artifact rejection system was used to reject trials (box and word) were administered to each participant. They began contaminated with eye blinks, eye movements or excessive with four items on the screen and this increased to six items, muscular activity. eight items and twelve items for the fourth trial. Reaction time and error rates were recorded. ERP Data Analysis Selected subtests of the Test of Everyday Attention (TEA) Waveforms were averaged by the Bio-Logic Brain Atlas (Robertson, Ward, Ridgeway, & Nimmo-Smith, 1996) were system. 30 single sweep waveforms were used to create an administered. The TEA has three parallel versions to account for average waveform for each task. ERP component wave forms practice effects from repeated administrations. The subtests used were determined through visual inspection. The most prominent in this study demonstrated high test-retest reliability. The Lottery peak occurring for all three electrode sites after the Nl-P2-N2 subtest is a ten minute test which requires participants to listen complex was chosen as the P300. Peak amplitude was measured for winning lottery numbers presented on audiotape. The test with respect to the mean of the 100 msec pre-stimulus baseline. authors report that this test loads on a sustained attention factor. Latency was measured with respect to stimulus onset and was The Map Search subtest was also administered. It requires defined as the time point of maximum positive amplitude. participants to search for symbols on a map. They are timed to see how many symbols they can locate in two minutes. This test Case Analyses and Results loads on a visual selective attention factor. Finally, the Auditory Elevator with Distraction subtest was administered. Participants Data were analyzed using visual inspection with graphing listened to beeping elevator sounds presented on audiotape and carried out as suggested by Krishef (1991). The primary researcher were required to count the floors in the presence of a distracting was blind to the identity and group membership of the tone. This subtest loaded on an auditory working memory factor participants during the data analysis phase of the study. and is a measure of selective attention and resistance to distraction. Attention and Working Memory Measures Because this study utilized alternate forms of the TEA, data Neurophysiological Measures were graphed as z scores to account for variability between forms. Event related potentials (ERPs) were recorded three times The tasks selected from the TEA were not as sensitive as during the baseline phase and once after each treatment phase expected. They were not challenging enough for a number of for a total of five recordings per participant. Brainstem auditory participants, who achieved ceiling level performance. The Lottery evoked potentials (BAEPs) and checkerboard pattern visual P100 task was the least sensitive of the three tasks utilized in this evoked potentials were recorded during the first test session study. Two participants did demonstrate improvement on TEA only. This was done to ensure that participants were free from tasks. Participant 2 showed improvement on the Auditory Elevator hearing or visual difficulties that could account for any with Distraction task after APT training (see Figure 1) and abnormalities noted in later component ERPs. improvement on the Map Search task after participation in the

16 Spring 2004 - The Journal of Cognitive Rehabilitation R E P O R T

Figure 1. Z scores for the auditory elevator with distraction task for Participant 2 for each test session.

AC condition delivered following APT. Baseline z scores for the three sites because site differences in amplitude, latency and the Map Search task ranged from -1.5 to -.7 with a post-treatment z relationship between amplitude and latency have been reported score of +l.5. Participant 4 demonstrated improvement on the (Ravden & Polich, 1999). Waveforms were examined for target Lottery subtest after participation in the APT condition (see and non-target stimuli. No participants demonstrated Figure 2). Improvement on the Auditory Elevator Task was noted abnormalities in their non-target evoked potentials (such as the during the AC condition delivered prior to APT (see Figure 3). presence of a large P300 waveform to non-target stimuli), This participant’s performance for the Map Search task showed suggesting no difficulty in discriminating between target and considerable variability in terms of z scores. The baseline was non-target stimuli. stable (z = 0) and improvements and decreases in performance All participants showed some change in ERPs. However, there were observed during the Active Control condition (z = -1.4 to was considerable variability in the changes observed and no +2). Considerable improvement was noted for the APT phase (z = clear pattern emerged. Decreased negativity of the N200 was +3 and +2.6). However, there was a drop to z = -.4 for the last test observed in three participants. Two participants showed changes session. This corresponds with the drop in scores also observed in N200 latency, however they were not consistent; there was for the Auditory Elevator with Distraction task. one increase and one decrease. Three participants showed The computerized working memory task proved to be a more decreased latency of the P300 component; no participants sensitive measure, with participants working at four levels of showed an increase in P300 latency. Three participants showed difficulty for the spatial and verbal tasks. All participants were increased P300 amplitude and two participants showed decreased able to achieve ceiling performance for the four item trials. The P300 amplitude for some tasks. These changes were all noted data were graphed by collapsing across levels of difficulty for all following participation in APT (see Table 2) with the exception of conditions that did not show ceiling effects or unstable baselines. a decrease in N200 latency observed for the auditory task Three of the four participants showed some improvement in following the AC condition for Participant 4. Actual amplitude error rates on this task following participation in APT (Participants and latency data are presented in Appendix A. 1, 2, 3) (see Figures 4, 5, 6). No participant showed improved performance on the working memory task after the Active Control Discussion condition without prior APT participation. In addressing the primary question posed by this study: is Neurophysiological Measures there a specific effect for the administration of attention training Evoked potential data was examined by graphing latency and (APT) above and beyond a supportive, adjustment oriented amplitude values for the P300 and N200 components. Data was approach, the data point to APT as being the efficacious graphed for each site individually (FZ, CZ, PZ). It was felt that ingredient in terms of producing cognitive change. With the inspection of amplitude and latency values for individual sites exception of Participant 4 showing improvement on the Auditory would be more sensitive than collapsing these values across all Elevator with Distraction task of the TEA following the AC

Spring 2004 - The Journal of Cognitive Rehabilitation 17 R E P O R T

Figure 2. Z scores for the lottery task for Participant 4 for each test session.

Figure 3. Z scores for the auditory elevator with distraction task for Participant 4 for each test session.

18 Spring 2004 - The Journal of Cognitive Rehabilitation Table 2. exercises is the active ingredient in producing change on R Summary of ERP Changes cognitive outcome measures and that there may be a differential E impact for attention retraining and psychotherapeutic support. However, due to the single case nature of this study, the small P # Auditory Auditory Visual Visual Task Task Task Task number of cases, and the relatively limited research bearing on O P300 N200 P300 N200 this question to date, replication and extension of these findings R will be important. T 1 Increased Decreased N2 Decreased P3 It is important to recognize that there were also some changes amplitude + negativity + amplitude noted in our data following participation in the AC condition. Decreased P3 This is consistent with the findings of others (Sohlberg et al., latency + 2000; Ruff & Neimann, 1990), and suggests that psychotherapeutic interventions can impact outcome variables 2 Decreased Decreased N2 Decreased P3 for some individuals. Clearly, the impact of psychotherapeutic P3 latency + negativity + amplitude interventions should be considered when assessing the efficacy Increased N2 latency of specific cognitive rehabilitation techniques. This has been recognized by the National Institute of Health in their consensus 3 No changes Increased P3 statement for the Rehabilitation of Persons with Traumatic Brain amplitude + Injury (1999). They stated that the findings from efficacy studies of rehabilitation for TBI have been limited by, among other things, 4 Decreased P3 Decreased N2 Increased P3 Decreased the unspecified effects of social contact. latency + latency +* amplitude + N2 neg. + The data also provided support for our second hypothesis, + indicates ERP differences believed to represent positive change. that neurophysiological variables would show the most change * only ERP changed noted following AC. after participation in APT. All participants showed some difference in ERPs that was believed to represent positive change. However, there was considerable variability and no clear pattern of change condition, and Participant 1 showing an improvement on the in ERPs emerged. For the two participants who received APT working memory task during the AC condition following APT, all prior to participation in the Active Control condition (Participants participants demonstrated improvement on the cognitive tasks 1 & 2), ERP changes were maintained, or in the case of Participant following participation in APT only. For the two participants who 1, actually became more pronounced. There was minimal change received APT prior to the AC condition, changes were maintained in ERP variables noted during participation in the AC condition during AC. For those who received AC first, changes were not for the participants who received the AC condition first. noted until attention retraining was initiated. The observed ERP changes were generally in the expected The findings provide support for improvement in performance direction. However, only one participant showed decreased N200 on tasks of working memory and attention following participation latency. Participant 2 showed increased N200 latency which was in attention retraining. These results are particularly positive in not predicted. Decreased negativity of the N200 component was light of the fact that improvements were demonstrated on observed in three participants suggesting that reduced effort untrained tasks that were quite different from the training tasks. was required to complete the task. Two participants showed The outcome measures employed provide stronger evidence of decreased latency of the P300 component. This is suggestive of generalization than studies that have used standard improved information processing speed. P300 latency is neuropsychological tests such as the PASAT, which is often considered to be a measure of stimulus classification speed or quite similar to training tasks and bears little similarity to functional information processing efficiency (Picton, 1992; Clark, 0-Hanlon, activities. TEA tasks were different from training tasks and Wright & Geffen, 1992). Polich (1998) describes P300 latency as a incorporated some elements of “real-life” demands. The sensitive temporal measure of neural activity underlying the computerized working memory tasks were an additional step processes of attention allocation and immediate memory. It has removed from training tasks in that they were delivered on been associated with superior cognitive performance in normal computer and introduced a memory component. subjects (Sturm, Willmes, Orgass & Hartje, 1997). The current findings add to the list of studies that have shown There were no hypotheses put forth with respect to attention retraining to be efficacious for individuals with TBI directionality of P300 amplitude changes. On the various tasks (Sohlberg & Mateer, 1987; Neimann, Ruff & Baser, 1990; Sturm, employed in this study, three participants showed increased P300 Willmes, Orgass & Hartje, 1997; Gray, Robertson, Pentland & amplitude and two showed decreased P300 amplitude. P300 Anderson, 1992; Gray & Robertson, 1989; Wilson & Robertson, amplitude changes are difficult to interpret because of the complex 1992; Ruff et al., 1994; Nag & Rao, 1999; Franzen & Harris, 1989; interactions of the influence of extraneous variables, such as Sohlberg et al., 2000; Palmese & Raskin, 2000; Cicerone et al., arousal level and task difficulty. Three participants in the current 2000). Our findings also contribute to a very small body of literature study demonstrated increased P300 amplitude values. Because a that has examined the differential impact of targeted cognitive reduction in P300 amplitude has been interpreted by many as an practice and more general, therapeutic or psychoeducational indication of reduced cognitive capacity, the increase may reflect approaches (Sohlberg et al., 2000; Ruff et al., 1989; Schmitter- improved attentional capacity. It may also reflect greater attention Edgecombe et al., 1995; Ruff & Neimann, 1990; Palmese & Raskin, given to the task stimuli, perhaps as a result of improved attention. 2000). The results suggest that the targeted practice on cognitive The decrease in amplitude observed in two participants can

Spring 2004 - The Journal of Cognitive Rehabilitation 19 R E P O R T

Figure 4. Participant 1: Errors for working memory task for all non-ceiling conditions combined.

Figure 5. Participant 2: Errors for working memory task for all non-ceiling conditions combined.

20 Spring 2004 - The Journal of Cognitive Rehabilitation R E P O R T

Figure 6. Participant 3: Errors for working memory task for all non-ceiling conditions combined. also be explained by examining the factors that affect the ampli- and 4). Neither of these participants was in coma, however both tude of the P300. Theoretically, a decrease in amplitude could demonstrated significant periods of post-traumatic amnesia (PTA). also reflect improved attentional abilities. When adequate atten- Another notable similarity of these two participants was the pres- tion is paid to a task, amplitude and certainty are believed to ence of depressive symptomatology. Clinical lore exists that sug- covary (Picton, 1992). Therefore more certainty of the correct gests that depression may be a contra-indicator for treatment response could be associated with decreased amplitude. In a because a depressed individual may not have the emotional re- similar vein, amplitude varies with task difficulty. A less difficult sources to benefit from a treatment program. Our results suggest task would produce a smaller amplitude P300. If participants’ otherwise. Involvement of individuals with depressive symp- attentional abilities have improved, the task may have become tomatology may provide an opportunity for behavioral activa- easier resulting in a smaller amplitude P300. Conversely, ampli- tion and distraction from or re-focusing of negative beliefs. How- tude also varies with arousal level and amount of attention paid ever, it should be noted that depression was measured in this to a task. Therefore, poor attention and/or effort could also result study using only the Rand depression inventory and was not in decreased amplitude. Examination of behavioral data for the assessed using specific diagnostic criteria (i.e., the DSM-IV). visual ERP task revealed that participants made few errors during Therefore, these findings may not hold true for individuals who all five administrations of this task. This weakens the notion that meet the DSM-IV criteria for a Major Depressive Disorder. reduced amplitude is evidence of reduced task difficulty or greater It has been suggested that treatment success is inversely certainty. The difference in directionality of amplitude changes related to time since injury (Franzen & Harris, 1993). However, observed across and within participants (Participant 1) weakens this finding may be somewhat spurious because numerous studies the evidence that bi-directional changes in amplitude reflect valid utilize patients who are still in a period of significant spontaneous, ERP changes related to the training tasks. Theory predicts that or natural, recovery. In the current study, participants ranged increased amplitude may reflect changes associated with im- from 15 months to 12 years post-injury, spanning a very wide proved attention but decreased amplitude is more likely related range of chronicity. Changes were shown by all participants, to changes in arousal level across test sessions. The findings of regardless of time since injury. This suggests that the inverse the current study add to a very small body of research that sug- relationship between time since injury and treatment success gests that ERP or EEG changes occur after cognitive rehabilita- may not hold for individuals who have reached a plateau in terms tion (Baribeau, Ethier & Braun, 1989; Raskin, 1996). of natural recovery. Examination of the current findings reveals that those partici- pants with the least severe initial injury showed change on the Limitations greatest number of behavioral and ERP measures (Participants 2 The present study addresses a number of issues that have

Spring 2004 - The Journal of Cognitive Rehabilitation 21 been raised as limitations of previous studies of the efficacy of designs: strategies for studying behavior change. New York: attention retraining (Park and Ingles, 2001). Repeated baseline Pergammon Press. R measurements and alternate forms were used to control for practice E effects and a control condition was employed. The heterogeneity Ben-Yishay, Y. & Diller, L. (1993). Cognitive remediation in traumatic inherent in the population was evaluated and not ignored in this brain injury: update and issues. Archives of Physical Medicine P study through its single case design methodology. The issue of Rehabilitation, 74, 204-213. O nonspecific, psychotherapeutic effects was systematically R examined. However, a number of limitations remain and must be Binder, L.M. (1986). Persistent symptoms after mild head injury: T discussed in considering the implications of the findings. A review of post-concussive syndrome. Journal of Clinical As noted previously, ceiling effects on the TEA tasks limited and Experimental , 8, 323-346. the ability to detect change in the different components of attention. Tasks were selected to represent sustained and selective Brooks, D.N., & McKinlay, W. (1987). Return to work within the attention, resistance to distraction, visual scanning and speed of first seven years of severe head injury. Brain Injury, 1, 5-15. processing. However, because only two participants showed improvements on these tasks it is not possible to address issues Campbell, K., Deacon-Elliot, D., & Proulx, G. (1986). relating to the responsivity of different attentional networks to Electrophysiological monitoring of closed head injury. I: Basic the training tasks. This is an important area for consideration as principles and techniques (the first of a three part series). some authors have suggested that degree of impairment in the Cognitive Rehabilitation, 4(5), 26-32. vigilance network may impact the outcome of training higher levels of attention (Sturm et al., 1997; Sohlberg et al, 2000). Carney, N., Chesnut, R.M., Maynard, H., Mann, N.C., Patterson, The duration of the rehabilitation program employed in the P., & Helfand, M. (1999). Effect of cognitive rehabilitation on current study was relatively short. Participants engaged in outcomes for persons with traumatic brain injury: A systematic rehabilitation activities for one hour sessions twice per week for review. Journal of Head Trauma Rehabilitation, 14(3), 277- six weeks. Stronger effects may have been observed for longer or 307. more intense training. Similarly, participants with more severe injuries may have benefited more from a longer duration of Chesnut, R.M., Carney, N., Maynard, H., Mann, N.C., Patterson, treatment. This study is also limited by a lack of follow-up data. P., Helfand, M. (1999). Summary report: Evidence for the This precludes the possibility of drawing any conclusions about effectiveness of rehabilitation for persons with traumatic brain the maintenance of treatment effects. Generalization to injury. Journal of Head Trauma Rehabilitation, 14(2), 176- improvement in real life functioning remains a challenge to 188. rehabilitation researchers. Our participants reported positive changes in their daily lives, however it is only possible to Cicerone, K.D., Dahlberg, C., Kalmar, K., Langenbahn, D.M., speculate about the role of the rehabilitation activities in these Malec, J.F., Bergquist, T.F., Felicetti, T., Giacino, J.T., Harley, improvements due to the influence of the individual’s J.P., Harrington, D.E., Herzog, J., Knelpp, S., Laatsch, L., & psychosocial environment outside of the rehabilitation program. Morse, P.A. (2000). Evidence-based cognitive rehabilitation: This study was conducted to attempt to determine the speci- recommendations for clinical practice. Archives of Physical ficity of the efficacy of attention retraining in producing improve- Medicine Rehabilitation, 81, 1596-1615. ments on tests of attention, working memory and neurophysi- ological measures, above and beyond a supportive, adjustment- Clark, C.R., O’Hanlon, A.P., Wright, A.P. & Geffen, G.M. (1992). oriented approach. Findings support the continued use of atten- Event-related potential measurement of deficits in information tion retraining activities for individuals with TBI who experience processing following moderate to severe closed head injury. attention deficits. Additional research is needed to elucidate spe- Brain Injury, 6(6), 509-520. cific relationships between patient characteristics, intensity and duration of administration of retraining activities, and treatment Deacon-Elliot, D., Campbell, K.B., Suffield, J.B.,& Proulx, G.B. approach. Future research utilizing ERPs and other functional Electrophysiogical monitoring of closed head injury. III. neuroimaging techniques may help to elucidate the neural pro- Cognitive evoked potentials (1987). Cognitive cesses underlying recovery after injury. Plasticity in the adult Rehabilitation, 5(3), 12-21. nervous system has potentially far-reaching implications for re- covery for various brain based disorders and injuries. Franzen, M.D., & Harris, C.V. (1993). Neuropsychological rehabilitation: application of a modified multiple baseline References design. Brain Injury, 7(6), 525-534.

Baribeau, J., Ethier, J., & Braun, C. (1989). A neurophysiological Gansler, D.A. & McCaffrey, R.J. (1991). Remediation of chronic assessment of selective attention before and after cognitive attention deficits in traumatic brain-injured patients. Archives remediation in patients with severe closed head injury. Journal of Clinical Neuropsychology, 6, 335-3 53. of Neuropsychological Rehabilitation, 3, 71-92. Gauggel, S. & Niemann, T. (1996). Evaluation of a short-term Barlow, D.W., & Hersen, M. (1984). Single case experimental computer assisted training programme for the remediation of attentional deficits after brain injury: a preliminary study.

22 Spring 2004 - The Journal of Cognitive Rehabilitation International Journal of Rehabilitation Research, 19, 229- with ADHD. Developmental Neuropsychology, 16(2), 273- R 23 9. 295. E Gevins, A. S., & Cutillo, B.A. (1971). Signals of Cognition. In A. KoIb, B. (1999). Synaptic plasticity and the organization of P Redmond (Ed.), Handbook of Clinical behavior after early and late brain injury. Canadian Journal O Electroencephalography and Clinical Neurophysiology (pp. of , 53(1), 62-75. R 335-379). Amsterdam: Elsevier. T Krishef, C.H. (1991). Fundamental Approaches to Single Subject Goransen, T. (1997). Evaluating the efficacy of cognitive Design and Analysis. Florida: Krieger Publishing rehabilitation: does attention training really work? Company. Unpublished manuscript. Luukinen, H., Viramo, P., Koski, K., Laippala, P., Kivela, S.L. (1999). Grafman, J. & Salazar, A. (1987). Methodological considerations Head-injuries and cognitive decline among older adults: a relevant to the comparison of recovery from penetrating and population based study. Neurology, 52(3), 557-562. closed head injuries. In H.S. Levin, J. Grafman, & H. M. Eisenberg (Eds.), Neurobehavioral recovery from head injury. Malec, J., Jones, R., Rao, N., & Stubbs, K. (1984). Video game New York: Oxford University Press. practice effects on sustained attention in patients with craniocerebral trauma. Cognitive Rehabilitation, 2, 18-23. Graves, R.E. (1991). The use of experimental techniques in clinical trials. In: E. Mohr and P. Brouwers (Eds.), Handbook of Middleton, D.K., Lambert, M.J., & Seggar, L.B. (1991). Clinical Trials, Amsterdam: Swets & Zeitlinger. Neuropsychological rehabilitation: microcomputer-assisted treatment of brain-injured adults. Perceptual and Motor Gray, J. & Robertson, I. (1989). Remediation of attentional Skills, 72, 527-530. difficulties following brain injury: three experimental single case studies. Brain Injury, 3(2), 163-170. Nag, S. & Rao, S.L. (1999). Remediation of attention deficits in head injury. Neurology India, 47, 32- 39. Gray, J., Robertson, I., Pentland, B., & Anderson, 5. (1992). Microcomputer-based attentional retraining after brain Neimann, H., Ruff, R., & Baser, C. (1990). Computer assisted damage: A randomised group controlled trial. attention retraining in head-injured individuals: A controlled Neuropsychological Rehabilitation, 2(2), 97-115. efficacy study of an outpatient program. Journal of Consulting and 58(6), 811-817. Gronwall, D.M.A. (1977). Paced Auditory Serial-Addition Tasks: A measure of recovery from concussion. Perceptual and NIH Consensus Statement Online (1999). Rehabilitation of Motor Skills, 44, 367-373. Persons With Traumatic Brain Injury. 16(1):1-41.

Hayes, S.C. & Leonhard, C. (1991). The role of the individual case Owens, A.M., Morris, R.G., Sahakian, B.J., Polkey, C.E., & Robbins, in clinical science and practice. In: M. Hersen, A. Kazdin, & A. T.W. (1996). Double dissociations of memory and executive Bellack (Eds.), The Clinical Psychology Handbook, 2nd Ed, functions in working memory tasks following frontal lobe New York: Pergamon Press. excisions, temporal lobe excisions or amygdalo~hippocampectomy in man. Brain, 119, 1597-1615. Heinze, H.J., Munte, T.F., Gobiet, W., Niemann, H., & Ruff, R.M. (1992). Parallel and serial visual search after closed head injury: Palmese, C.A., & Raskin, S.A. (2000). The rehabilitation of electrophysiological evidence for perceptual dysfunctions. attention in individuals with mild traumatic brain injury, using Neuropsychologia, 30(6), 495-514. the APT-TI programme. Brain Injury, 14(6), 535- 548.

Kazdin, A.E., (1984). Statistical analysis for single-case Papincolaou, A.C., Levin, H.S., Eisenberg, H.M., Moore, B.D., experimental designs. In: D.H. Barlow & M. Hersen (Eds.), Goethe, K.E., & High, W.M. Jr. (1984). Evoked potential Single Case Experimental Designs, nd Ed. New York: correlates of posttraumatic amnesia after closed head injury. Pergamon Press. 2 Neurosurgery, 14 (6), 676-678.

Keren,O., Ben-Dror, S., Stern, M.J., Goldberg, G. & Groswasser, Z. Park, N.W. & Ingles, J.L. (2001). Effectiveness of attention (1998). Event-related potentials as an index of cognitive rehabilitation after an acquired brain injury: A meta-analysis. function during recovery from severe closed head injury. Neuropsychology, 15(2), 199-210. Journal of Head Trauma Rehabilitation, 13(3), 15-30. Park, N.W., Proulx, G.B., & Towers, W.M. (1999). Evaluation of Kerns, K.A. (1997). Working Memory Task. University of Victoria. attention training. Neuropsychological Rehabilitation, 9(2), 135-154. Kerns,K.A., Eso, K., & Thomsen, J. (1999). Investigation of a direct intervention for improving attention in young children

Spring 2004 - The Journal of Cognitive Rehabilitation 23 Picton, T. W. (1992). The P300 wave of the human evoked potential. Stone, S. & Raskin, S. (1996). Training prospective memory in an Journal of Clinical Neurophysiology, 9(4), 456-479. individual with anoxic brain damage. Poster session R presented at International Neuropsychological Society. E Pouch, J. (1998). P300 clinical utility and control of variability. Chicago, IL. Journal of Clinical Neurophysiology, 15(1), 14-33. P Sturm, W., Willmes, K., & Orgass, B. (1997). Do specific attention O Ponsford, J. & Kinsella, G. (1988). Evaluation of a remedial program deficits need specific retraining? Neuropsychological R for attentional deficits following closed-head injury. Journal Rehabilitation, 7(2), 81-103. T of Clinical and Experimental Neuropsychology, 10(6), 693- 708. Wilson, B.A. (1993). How do we know that rehabilitation works? Neuropsychological Rehabilitation, 3(1), 1-4. Ponsford, J., & Kinsella, G. (1991). The use of a rating scale of attentional behaviour. Neuropsychological Rehabilitation, Wilson, C., & Robertson, I.H. (1992). A home-based intervention 1 (4), 241-257. for attentional slips during reading following head injury: a single case study. Neuropsychological Rehabilitation, 2(3), Raskin, S. (1996). P300 as a measure of brain reorganization 193-205. following cognitive rehabilitation Poster presented at Cognitive Neuroscience Society. San Francisco, CA.

Robertson, I.H., Ward, T., Ridgeway, V., & Nimmo-Smith, I. (1996). The structure of normal human attention: the Test of Everyday Attention. Journal of the International Neuropsychological society, 2, 525-534.

Ruff R., Baser, C.A., Johnston, J.W., Marshall, L.F., Klauber, S.K., Klauber, M.R., & Minteer, M. (1989). Neuropsychological rehabilitation: An experimental study with head-injured patients. Journal of Head Trauma Rehabilitation 4(3), 20- 36.

Ruff R. & Niemann, H. (1990). Cognitive rehabilitation versus day treatment in head-injured adults: is there an impact on emotional and psychosocial adjustment? Brain Injury, 4(4), 339-347.

Rugg, M.D., Cowan, C.P., Nagy, M.E., Milner, A.D., Jacobson, I. & Brooks, D.N. (1988). Event-related potentials from closed head injury patients in an auditory “oddball” task: evidence of dysfunction in stimulus categorization. Journal of Neurology, Neurosurgery, and Psychiatry, 51, 691-698.

Schmitter-Edgecombe, M., Fahy, J.F., Whelan, J.P. & Long, C.J. (1995). Memory remediation after severe closed head injury: notebook training versus supportive therapy. Journal of Consulting and Clinical Psychology, 63 (3), 484-489.

Sohlberg, M.M. & Mateer, C.A. (1987). Effectiveness of an attention training program. Journal of Clinical and Experimental Neuropsychology, 9(2), 117-130.

Sohlberg, M.M. & Mateer, C.A. (1989). Introduction to Cognitive Rehabilitation. theory and practice. New York: Guilford Press.

Sohlberg, M.M., McLaughlin, K.A., Pavese, A., Heidrich, A., & Posner, M.I. (2000). Evaluation of attention process training and brain injury education in persons with acquired brain injury. Journal of Clinical and Experimental Neuropsychology, 22(5), 656-676.

24 Spring 2004 - The Journal of Cognitive Rehabilitation Appendix A: ERP Changes by Participant R E Participant 1 P Auditory Task: O R This participant demonstrated an increase in amplitude of the P300 after APT. The observed increased amplitude continued after the T AC condition. Baseline amplitudes for the P300 were 5.2 mV, 1.7 mV, 1.82 mV. Amplitude after APT only was 6.61 mV and after APT and AC was 7.72 mV. P300 latency also demonstrated a decrease; however, it is difficult to interpret this data due to an unstable baseline. Participant 1 demonstrated decreased negativity of the N200 for the auditory task after APT that continued after participation in the AC condition. Baseline amplitudes for the N2 were -1.69 mV, -2.11 mV, -2.0 mV. Amplitude after APT only was -1.0 mV and after APT and the AC condition was 1.54 mV. Baseline latencies for N2 were not stable.

Visual Task:

Participant 1 did not show corresponding changes for the visual task in the P300 or N200 components, however, some changes were observed. A decrease in P300 amplitude was observed for this task and there was a decrease in P300 latency after APT. Baseline amplitudes for the P300 were 10.77 mV, 18.59 mV, 11.68 mV. Amplitude after APT only was 7.59 mV and after APT and AC was 4.25 mV. Latencies for baseline measurements were 436 msec, 424 msec, 604 msec. Latency after APT only was 316 msec and after the AC condition was 300 msec. Reaction time (RT) for this task also showed a decrease from baseline. Baseline reaction times were 366.5 msec, 358 msec, 391 msec. After APT, RT for the visual task was 341 msec. RT was 341.5 after the AC condition. N2 data could not be interpreted for this task because of unstable baselines.

Participant 2

Auditory Task:

This participant showed two ERP changes for the auditory task after participation in APT. There was a decrease in the negativity of the N200 and an increase in N200 latency following APT. There was no change in amplitude of the P300. There were changes in P300 latency but they were judged to be insignificant. Baseline amplitudes for the N200 were-2.91 mV, -3.55 mV, -3.36 mV. After participation in APT, N200 amplitude was .47 mV. After participation in the AC condition, amplitude was -.3 mV. Baseline latencies of the N200 were 256 msec, 216 msec, 220 msec. After participation in APT, latency was 288 msec, and was 280 msec after participation in the AC condition.

Visual Task:

Participant 2 showed decreased P300 amplitude after participation in APT. Baseline amplitudes for the P300 were 9.34 mV, 10.49 mV, 9.99 mV. After participation in APT, P300 amplitude was 7.23 mV. After participation in the AC condition, amplitude was 7.36 mV. There was variability in baseline data for P300 latency for the visual task. There were no observed changes in the N200 component.

Participant 3

Auditory Task:

This participant showed no changes in either the P300 or the N200 components after participation in either condition for the auditory task.

Visual Task:

Participant 3 showed an increase in P300 amplitude following APT. Only two baseline measurements were available for this partici- pant due to a poor quality recording. Baseline measures for P300 amplitude were 6.97 mV and 9.25 mV. After participation in the AC condition, amplitude was 8.07 mV. After APT amplitude was 20.53 mV. P300 latency was variable.

Spring 2004 - The Journal of Cognitive Rehabilitation 25 Participant 4 R Auditory Task: E P Participant 4 showed decreased latency for the P300 following APT and decreased latency for the N200 following AC. This partici- pant missed one baseline session; therefore, there are only two data points for baseline data. Baseline measures for latency of the P300 O were 616 msec and 392 msec. After participation in the AC condition, P3 latency was 368 msec. After APT, latency was 264 msec. Baseline R measures for latency of the N200 were 408 msec, 252 msec. After the AC condition, latency was 204 msec. After APT it was 200 msec. No T changes in amplitude of either component were observed.

Visual Task:

This participant showed an increase in P300 amplitude and a decrease in negativity of the N200 for the visual task with no corresponding changes in latency of either component following APT. Baseline measures for P300 amplitude were 4.82 mV, 7.04 mV. After the AC condition, amplitude was 3.83 mV and after APT amplitude was 11.27 mV. Baseline amplitude measures for the N200 were -4.82 mV and -6.33 mV. After participation in the AC condition, amplitude was -6.32 mV. After APT participation, amplitude was -1.78 mV.

26 Spring 2004 - The Journal of Cognitive Rehabilitation